Overview
- The Wall Street Journal reported on June 11 that OpenAI has discussed steep reductions to the per‑token fees it charges business customers, and people familiar with the talks say no final decision has been made.
- The move would target 'tokens'—the metered units used to bill API and enterprise use—rather than flat consumer subscription fees, so large customers who pay by usage would see the biggest price change.
- Analysts warn deep token cuts would squeeze margins because both OpenAI and Anthropic run on costly GPU compute and large cloud or colocation deals that drive high unit costs.
- Anthropic’s recent funding surge and revenue growth, including a reported ~$65 billion round valuing it near $965 billion, has given it market momentum that is pressuring OpenAI to consider defensive pricing.
- Lower‑cost Chinese and open‑source inference providers are already undercutting Western prices by multiples, and corporate pushback over exploding AI bills has firms rationing usage and demanding cheaper options.